There have been many reactions (and overreactions) to the iOS location tracking fiasco. I won’t try to explicate the ramifications for privacy and security. But the reaction that resonated most for me was probably this one by Marco Arment.
Nevertheless, on the day that the issue went mainstream (it had been known by forensics experts for months), I downloaded the iPhone Tracker application and ran it on my own data. And I noticed something strange, which I think adds credence to the theory that this is merely a cache that is not being properly culled (a theory espoused by John Gruber). Now it’s possible that this observation is somehow unique to my circumstances but I doubt it.
I usually travel a fair amount, but my travel has been severely curtailed since the release of iOS 4 due to the birth of our third child on August 1. Thus, I have taken only a handful of trips, all domestic, since then. When I opened iPhone Tracker, there, on a map of the US, was a plot of everywhere that my iPhone had connected to the cellular network since the middle of last summer. I saw my drive through Texas in a rental car with my son, airplane connections in Charlotte, Atlanta, and Chicago, driving trips to Washington, DC and Frederick, MD, and visits to Miami and Los Angeles. It was a little creepy, especially as I didn’t remember that I had connected in Chicago on my way to LA and probably would have sworn to you that I hadn’t been through the Chicago airport in the last year.
But here’s the strange thing: I have spent the vast majority of my time these last 10 months within 15-20 miles of my home in Blacksburg. While every cell tower in Blacksburg and the surrounding vicinity was marked, none of them were marked except in the last week. (The iPhone Tracker app lets you step through the traces week-by-week.) Clearly, older visits to a particular cell tower (or in a particular area) are being culled from the file.
Thus, the phone (and, by extension, iTunes on the computer) appears to maintain an accurate record of everywhere that it has been in the last 10+ months, but it only appears to keep track of the most recent visits to each location. To me, this points directly at a poorly designed or implemented culling strategy. When “tracking points” are added to the file, my guess is that if the number of “similar” points exceeds some threshold, then the oldest similar points are removed. I haven’t investigated closely enough to try to figure out the exact culling strategy, and I probably won’t. I would be interested to know whether or not other people’s iPhone traces exhibit similar patterns.